Boolean combinations of conditions on vertex visibility

Jan 21, 2014 at 2:10 AM
If I wanted to see only the vertices whose clustering coefficient is greater than 0.9 OR betweenness centrality is greater than 200, how would I do that? Possible?
Jan 21, 2014 at 3:03 AM
Edited Jan 21, 2014 at 3:03 AM
You may need to create an =IF() formula in the vertex visibility column in the Vertices worksheet that references those values. Let the IF resolve to zero if it fails the test.

Jan 21, 2014 at 7:51 PM
I found being able to do this was extremely valuable on a previous project, using a different network visualization/analysis package. Unfortunately, as soon as the company realized how many people were using the feature, the package went from free to very
expensive. It operated on attributes, e.g., sex = female AND age > 16. The feature was implemented in such a way that learning to use it was difficult, and it took some amount of thought to use it cleverly, but it put the package on a whole different plane.
What a luxury it would be to have it available in NodeXL! I realize it could be done with user macros, but to make it general would be a fairly hefty programming effort. Peter
Jan 21, 2014 at 7:59 PM
Filtering vertices and edges with complex, compound filters is simple in NodeXL.

From the Graph Pane, select Dynamic Filters. This dialog presents all vertex and edge attributes in the form of double headed sliders. You may set any number of filters for betweenness or clustering coefficient in this dialog and only the edges and vertices that pass this filter will be displayed.

This task can also be performed more manually using an =IF() formula in the "Visibility" column of the Edges and Vertices worksheets. If this column contains a zero for an edge or a vertex it will not be displayed or considered in calculations when they are next performed.

An =IF() statement can be nested, allowing for multiple conditions to be tested.

Jan 21, 2014 at 8:21 PM
Marc, thank you for reminding that one can just enter Excel =If() expressions and do 'Manual Fill" -- I should have thought of this, there is an example in the book! Entering complex expressions, especially nested ones, can be tedious, but they can be prepared in a text editor and pasted in the cell. Thank you also for clarifying that a zero in the visibility column not only hides the data but also excludes it from further calculations, unlike Dynamic Filters. Also, Dynamic Filters cannot deal with categorial properties, like Country of origin, unless I'm missing something. This would be a nice addition.
Tue at 6:34 PM
Dear Marc Smith,

I will be again teaching a unit on SNA in my computer for non-majors class, and I've just discovered that many features of NodeXL that used to be standard base-line install are now in NodeXL Pro and cost money. I am completely unprepared for this, and don't have the funds. I can pay for my personal copy, but do you think it would be possible to waive the fee for the 10 students in my class. We can solemnly swear that the installations will be removed once the semester is over, Dec. 9, 2016.

Many thanks!

Alexander Nakhimovsky
Director, Linguistics Program
Director, Program on Security and Democracy
Colgate University Hamilton NY 13346
t. +1 315 228 7586